Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!samsung!munnari.oz.au!lee
From: lee@munnari.oz.au (Lee Naish)
Newsgroups: comp.lang.prolog
Subject: Re: Fun. vs. Logic
Message-ID: <2788@munnari.oz.au>
Date: 23 Nov 89 09:21:25 GMT
References: <11500018@hpldola.HP.COM> <11500019@hpldola.HP.COM> <DJK.89Nov23184835@munginya.cs.mu.OZ.AU>
Sender: news@cs.mu.oz.au
Reply-To: lee@munmurra.UUCP (Lee Naish)
Organization: Comp Sci, University of Melbourne
Lines: 35

In article <DJK.89Nov23184835@munginya.cs.mu.OZ.AU> djk@cs.mu.Oz.AU writes:
>...
Astute readers may have noticed some overlap with my posting - sorry if
it bored you, but I couldn't be bothered cancelling my article and
rewording it to avoid overlap.  Really clever readers would have noticed
that David and I contradicted each other:
>
>]     ?- not person(X).
>
>Your query fails immediately, because X only occurs once.

The reason for the disagreement is a useful but unfortunately confusing
feature of NU-Prolog, namely implicit quantification.  NU-Prolog allows
you to write explicit quantifiers as follows:

?- all X not p(X).

This query fails, since there is an X such that p(X).  It turns out that
some code ends up looking like this:

	all [X1,X2,X3,X4] not p(RelevantVar, X1, X2, X3, X4)

which looks much nicer if you allow implicit quantifiers:

	not p(RelevantVar, _, _, _, _)

Variables can be implicitly quantified if they occur only once in a
clause (and there is a style checker which warns you if such variables
are not prefixed with an underscore).  Last time I had anything to do
with the implementation, this implicit quantification was not done with
queries at the top level of the system.  This has changed (I guess I
should have known), so X in the original query is treated as universally
quantified outside the negation (or existentially inside).

	lee